json-numpy provides lossless and quick JSON encoding/decoding for NumPy arrays and scalars.
json-numpy follows Semantic Versioning.
json-numpy can be installed using pip:
$ pip install json-numpy
For a quick start, json_numpy can be used as a simple drop-in replacement of the built-in json module.
The dump(), load(), dumps() and loads() methods are implemented by wrapping the original methods and replacing the default encoder and decoder.
More information on the usage can be found in the json module's documentation.
import numpy as np
import json_numpy
arr = np.array([0, 1, 2])
encoded_arr_str = json_numpy.dumps(arr)
decoded_arr = json_numpy.loads(encoded_arr_str)Another way of using json_numpy is to explicitly use the provided encoder and decoder functions in conjunction with the json module.
import json
import numpy as np
from json_numpy import default, object_hook
arr = np.array([0, 1, 2])
encoded_arr_str = json.dumps(arr, default=default)
decoded_arr = json.loads(encoded_arr_str, object_hook=object_hook)Finally, the last way of using json_numpy is by monkey patching the json module after importing it first:
import json
import numpy as np
import json_numpy
json_numpy.patch()
arr = np.array([0, 1, 2])
encoded_arr_str = json.dumps(arr)
decoded_arr = json.loads(encoded_arr_str)This method can be used to change the behavior of a module depending on the json module without editing its code.
The simplest way to run tests is:
$ python -m unittest
As a more robust alternative, you can install tox to automatically test across the supported python versions, then run:
$ tox
Please report any bugs or enhancement ideas using the issue tracker.
json-numpy is licensed under the terms of the MIT License (see LICENSE.txt for more information).